System and Method of Searching a Corpus

1. A search method comprising: creating a list of candidate probe words; for each candidate probe word, counting a number of item descriptions that contain the candidate probe word; choosing, with a computer, q i probe words from the candidate probe words whose word count is closest to |D i |/(q i +1), where D i is a set of remaining item descriptions after a previous repetition, and q i represents a number of probe words presented for user selection for the ith repetition and is an integer greater than 1; presenting the q i probe words for selection by the user; receiving a selection of one of the q i probe words by the user; pruning the list of candidate probe words to eliminate item descriptions that include non-selected ones of the q i probe words to create a pruned list of candidate probe words; choosing q i+1 probe words from the pruned list of candidate probe words that is closest to |D i+1 |/(q i+1 +1), wherein D i+1 is a set of remaining item descriptions after the ith repetition, q i+1 represents a number of probe words presented for user selection for the i+1 repetition and is an integer greater than 1, and the pruned list of candidate probe words includes the selected one of the q i probe words; and presenting the q i+1 probe words for selection by the user.

2. The method recited in claim 1 , further including creating a new list of candidate probe words from the remaining item descriptions after the ith repetition wherein creating the new list of candidate probe words is performed after pruning the list of candidate probe words.

3. The method recited in claim 1 , further including presenting a final list of items for user selection.

4. The method recited in claim 1 , wherein: the word count of the q i probe words most equally divides the number of remaining item descriptions after the previous repetition into q i +1 subgroups; and the q i +1 subgroups are disjoint subgroups.

5. The method recited in claim 1 , wherein the item descriptions are capable of being changed with each repetition.

6. The method recited in claim 1 , wherein creating the list of candidate probe words further includes eliminating any noise words from the list of candidate probe words and normalizing the list of candidate probe words.

7. A non-transitory computer readable storage medium having computer readable program instructions stored thereon for causing a computer system to perform instructions, the instructions comprising the steps of: creating a list of candidate probe words; for each candidate probe word, counting a number of application descriptions that contain the candidate probe word; choosing q probe words from the candidate probe words whose word count is closest to |D i |/(q+1), where D i is a set of remaining application descriptions for an ith repetition, and q is a number of probe words presented for user selection and is an integer greater than 1; presenting the q probe words for user selection; pruning the list of candidate probe words to eliminate application descriptions that include non-selected ones of the q probe words to create a pruned list of candidate probe words, wherein the pruned list of candidate probe words includes a probe word selected by the user from the q probe words; and repeating counting the number of application descriptions, choosing q probe words from the pruned list of candidate probe words, presenting the q probe words for selection and pruning the list of probe words until a final list of applications remain.

8. The non-transitory computer readable storage medium having computer readable program instructions recited in claim 7 , the instructions further comprising presenting the final list of remaining applications.

9. The non-transitory computer readable storage medium having computer readable program instructions recited in claim 7 , the instructions further comprising creating a new list of candidate probe words from the remaining item descriptions wherein the step of creating a new list of candidate probe words is performed after the pruning step.

10. The non-transitory computer readable storage medium having computer readable program instructions recited in claim 7 , the instructions further comprising presenting the final list of applications for user selection.

11. The non-transitory computer readable storage medium having computer readable program instructions recited in claim 7 , wherein q+1 subgroups are disjoint subgroups.

12. The non-transitory computer readable storage medium having computer readable program instructions recited in claim 8 , wherein the application descriptions are capable of being changed with each repetition.

13. The non-transitory computer readable storage medium having computer readable program instructions recited in claim 7 , wherein creating a list of candidate probe words further includes the steps of eliminating any noise words from the list of candidate probe words and normalizing the list of candidate probe words.

14. A computer hardware system for generating probe words comprising: a candidate word component for generating a list of candidate probe words; a description word counter that counts a number of item descriptions that contain the candidate probe words for each candidate probe word; a selection component for choosing q probe words whose word count is closest to |D i |/(q+1), where D i is a set of remaining item descriptions for an ith repetition, and q is a number of probe words presented for user selection and is an integer greater than 1; a presentation component for presenting a list of q probe words for selection by the user; a pruning component for eliminating non-selected ones of the q probe words and item descriptions associated with the non-selected ones of the q probe words to create a pruned list of candidate probe words, wherein the pruned list of candidate probe words includes a probe word selected by the user from the q probe words; and the candidate word component for generating a new list of candidate probe words from the pruned list of candidate probe words; wherein the candidate word component, description word counter, selection component, presentation component, and pruning component are executed by a processor communicatively coupled to a memory.

15. The computer hardware system for generating probe words recited in claim 14 , wherein the candidate word component can optionally create a new list of candidate probe words from the remaining item descriptions, wherein the step of creating a new list of candidate probe words is performed after the pruning component eliminates previously chosen probe words and items associated with the probe words that were not selected.

16. The computer hardware system for generating probe words recited in claim 14 , wherein the presentation component further presents a final list of item descriptions for user selection.

17. The computer hardware system for generating probe words recited in claim 14 , wherein q subgroups are disjoint subgroups.

18. The computer hardware system for generating probe words recited in claim 14 , wherein the item descriptions are capable of being changed with each repetition.